Page 63 - IJAMD-2-3

International Journal of AI for
Materials and Design Intelligent interactive textile in healthcare

wall panels that were knitted with POFs and embedded The integration of illuminative fabric into an AI-driven
with RGB LEDs to enable illumination. These fibers guide gesture recognition system creates a novel multi-sensory
emitted light across the textile surface, emitting the LED and contactless interaction platform, specifically aimed at
output and transforming it into soft, color-based visual promoting engagement, exercise, and therapy in healthcare
feedback. In the first prototype (Version 1) shown in settings. From a design and engineering perspective, the
Figure 4D, the response system was relatively simple: knitted POF fabric served both esthetic and functional
Classified gesture data were sent from the Raspberry goals by emitting programmable RGB light in response
Pi to a basic relay switch connected to the RGB LEDs. to validated gestures. This visual feedback mechanism –
This setup allowed the system to turn different colored subtle, localized, and dynamic – provided a soft and non-
light channels on or off based on the detected gesture or intrusive way for users, particularly elderly individuals,
body posture, directly illuminating the textile surface in to interact with therapy-driven content. The iterative
specific colors corresponding to each gesture and posture transition from the first prototype, which relied on basic
type. In the advanced version (Version 2) shown in relay control, to an advanced custom-made PCB solution
Figure 4F, a more modular and intelligent infrastructure demonstrates the scalability of the design. The upgraded
was implemented. The system included a self-developed use of an ESP32 controller introduced advanced data
PCB built around an ESP32 microcontroller, which handling, reduced latency, and allowed greater precision
decodes serial input from the single-board computer and over RGB transitions, thereby enabling alignment with
distributes PWM signals to the RGB LEDs. The advanced the AI model and optimizing the best color combinations
system also supports potential Internet of Things across different panel designs and illumination patterns.
applications, while simplifying hardware integration and This technical refinement enhanced the responsivity and
improving system stability and responsiveness. The RGB fluidity of the textile interface, making the system feel
LEDs, embedded along the edges of the textile panel and more intuitive and responsive to natural body movement.
optically coupled through POFs, share a common +5V Furthermore, the edge-mounted RGB LEDs, optically
power source. Control signals (R, G, B) were adjusted coupled to the textile through POFs, enabled a contactless
through the PCB using serial data inputs processed by yet emotionally resonant interaction method, particularly
a state machine methodology, ensuring synchronized valuable in post-pandemic healthcare environments
responses with gesture inputs. A unified 12V power focused on hygiene and psychological comfort. The
supply supported all electronic components within the textile’s ability to visually signify correct gestures with
system, delivering power directly to the custom PCB soft illuminations empowers elderly users through instant
and to a voltage converter that steps down the current feedback, supporting self-guided physical and cognitive
to 5V for the single-board computer. This centralized rehabilitation. Overall, this integration not only reinforces
power configuration ensures a compact, efficient, and the practicality of smart textiles for ambient healthcare but
easily maintainable setup, well-suited for integration also demonstrates how co-designed, AI-enhanced systems
in healthcare environments. Figure 4E and G show the can drive inclusive and meaningful experiences in spatial
color differences corresponding to various gestures and interaction, particularly for vulnerable populations.
body movements in the fabric wall panels for Version
1 and Version 2, respectively. In Version 1, the left wall 4.4. System validity and privacy considerations
panel was designed to respond to shoulder movements. To support real-world deployment in healthcare contexts,
The illumination changed to blue, pink, and a dynamic the design and implementation of the gesture recognition
“jumping” color effect when the system recognized the system deliberately accounted for environmental factors
gestures “hands up,” “open arms,” and the WTSDHC such as lighting conditions and gesture visibility. While
slogan gesture “My Health, My Say!” performed by no quantitative accuracy metrics were recorded at this
pointing to oneself, respectively. The middle panel was stage, system latency was monitored through built-in
dedicated to hand gesture recognition, with color changes software timers, which counted frame-processing loops
of yellow, blue, and pink corresponding to the gestures based on timestamp differences. The average processing
“good,” “OK,” and “love.” The right panel was designed for rate was approximately 5 – 7 times per second, confirming
head and neck movements; when the user turned their operational responsiveness suitable for real-time
head “up,” “down,” “left,” or “right,” the textile illumination interaction. During participatory co-design workshops,
changed accordingly to blue, yellow, green, and pink. the viewing angle and lighting conditions were thoroughly
Figure S3 illustrates the infographic interaction with discussed with stakeholders. Based on this input, the optimal
the illuminative fabric wall panels through (a) shoulder, interaction distance was standardized at approximately 1
(b) hand, and (c) head movements, respectively. meter, with lighting conditions kept consistent to ensure

Volume 2 Issue 3 (2025) 57 doi: 10.36922/IJAMD025170013